The simplest is to use
[ux,ia]=unique(x);
vi=interp1(ux,y(ia),v);
This will use the first of the y values for each unique x; you could get more sophisticated and find those values and average them or otherwise set the interpolant value at that value.
Alternatively, I have on occasion found the duplicate locations and adjusted their positions by ux+/-2*eps(ux) to make them just far enough apart that interp1 doesn't complain. This is useful if there is a real discontinuity in the interpolant at the point and it is desired to model it as well.
